matrix - Eigenvalues of huge matrices -
i have 2 50,000 x 50,000 symmetric real matrices , b (double precision) non-sparse (55% non-zero). b positive-definite.
i have generalized eigenproblem: v = ยต b v
i need find 3 - 4 of smallest algebraic eigenvalues (and if possible associated eigenvectors).
do have still options left compute them on average computer 12 gb of ram?
any suggestions can either try without effort or work. thanks!
your options use out-of-core solvers.
here scalapack reference: http://www.netlib.org/scalapack/prototype/readme.outofcore
but have time this, isn't it? getting few eignenvalues can render problem feasible in reasonable amount of time, recommend estimate needed time before starting code.
Comments
Post a Comment